Pore structure and mechanical properties in freeze cast porous Si3N4 composites using polyacrylamide as an addition agent
Feng Ye,Jingyi Zhang,Haijiao Zhang,Limeng Liu
DOI: https://doi.org/10.1016/j.jallcom.2010.07.020
IF: 6.2
2010-01-01
Journal of Alloys and Compounds
Abstract:Research highlights ▶ Porous BAS/Si 3 N 4 composites with unidirectional aligned channels were fabricated by freeze casting using liquid N 2 as refrigerant. ▶ Increasing PAM content in the slurries strongly affect the pore structure, α → β -Si 3 N 4 transformation and mechanical properties. Abstract Porous barium aluminosilicate (BAS)/Si 3 N 4 composites were prepared by freeze casting process using liquid N 2 as refrigerant. The pore morphologies, α → β -Si 3 N 4 phase transformation, and mechanical properties of the resultant porous ceramics were strongly affected by the addition of polyacrylamide (PAM) into the slurries. The incorporation of 2 wt.%PAM changed the ice crystals structure from lamellar with dendrites into featureless columnar ones, the feature of which was duplicated by the pores in the sintered ceramics. The lamellar pores with dendrites could favor the α → β -Si 3 N 4 phase transformation, but were detrimental to the mechanical properties. The sample with 1 wt.%PAM exhibited the highest strength and toughness in the resultant porous composites due to its columnar pore structure with less dendritic pores and higher aspect ratio of β -Si 3 N 4 grains. A relative dense layer was formed at the bottom of the samples with PAM and its thickness increased with increasing PAM content. Keywords Porous Si 3 N 4 ceramics Polyacrylamide Freeze casting Microstructure Mechanical properties 1 Introduction Advanced porous ceramics with high porosity and large surface area were widely used as filters, bio-prosthetic implants, catalyst supports and lightweight structural materials [1,2] . Porous silicon nitride is a promising candidate owing to their excellent mechanical and thermal–mechanical properties [3,4] . A number of manufacturing techniques, namely adding fugitive substance [5] , partial sintering by using less sintering aid [6,7] , partial hot-pressing [8] have been developed for the production of porous silicon nitride ceramics. Recently, the freeze casting process has been developed to produce highly porous ceramics, because it can produce interconnected pore channels in a controllable manner, which offers superior mechanical properties and functions [9] . Freeze casting has already been used for fabricating a wide variety of ceramic materials [10–13] . However, till now, the reports on the microstructural control and mechanical properties of porous silicon nitride ceramics fabricated by freeze casting are very scarce. Polyacrylamide (PAM) is a water soluble polymer of acrylamide monomers. It can be used in a variety of industries including food, water treatment and agricultural industries. Barium aluminosilicate (BAS) has been proved to be an excellent sintering additive to promote the α - to β -Si 3 N 4 transformation and the growth of elongated β -Si 3 N 4 grains [14] . The current work is concerned with the microstructure control of porous BAS/Si 3 N 4 composites by changing the PAM content. 2 Experimental procedures The materials used in this study were BAS/Si 3 N 4 composites with 10 wt.%BAS. Starting materials were Si 3 N 4 (supplied by Shanghai Junyu Ceramic Co., Ltd., China, α -Si 3 N 4 content >95%, average particle size is 0.5–1.0 μm) and BAS glass–ceramic. The BAS glass–ceramic powder was synthesized through the hydrolysis of alkoxides [15] . First, BAS and Si 3 N 4 powders were well dispersed in the distilled water with 0.3 wt.% polyacrylic acid (PAA) (based on powders) as dispersant and different amount of PAM (Shanghai Chemical Reagent Corp., China), and then ball-milled for 20 h to obtain stable slurries. In order to investigate the effect of PAM concentration on the microstructure of the samples, the solid loadings of the slurries were fixed to 40 vol.% and 0, 1 and 2 wt.%PAM (based on water) were dissolved in the water. The prepared slurries were then poured into polyethylene molds ( ϕ 70 mm × 15 mm) using a brass bottom plate immersed in liquid N 2 . The top of the container was open so that the upper surface of the slurry would expose to the atmosphere at room temperature. Immediately after casting, the ice crystals grow unidirectionally from the bottom to the upper surface of the slurry. After the frozen BAS/Si 3 N 4 samples were completely freeze-dried under vacuum for 2 days, they were carefully placed into a graphite crucible with a silicon nitride-based powder bed and sintered in a graphite resistance furnace at 1800 °C for 2 h under a 0.1 MPa nitrogen atmosphere. The heating and cooling rates were both 5 °C/min. The densities of the samples were calculated from the dimensions and weight of the samples. Pore size distribution was measured by mercury porosimetry (Model Autopore III 9420, Micrometics Co., USA). Flexural strength and fracture toughness were measured in air at room temperature. All flexural bars were machined with the tensile surface perpendicular to the freezing direction. Flexural strength measurements were performed on bar specimens (3 mm × 4 mm × 36 mm) using a three-point bend fixture with a span of 30 mm. Fracture toughness measurements were performed on single-edge-notch beam specimens (2 mm × 4 mm × 30 mm) with a span of 16 mm, and a half-thickness notch was made using a 0.1 mm thick diamond wafering blade. At least, six specimens were tested for each test condition. Crystalline phases of the sintered Si 3 N 4 /BAS composites were characterized by X-ray diffraction (XRD). Quantitative analysis of α - and β -Si 3 N 4 phase content was determined by comparing the peak intensity ratios [16] . The microstructure of porous Si 3 N 4 samples was characterized by scanning electron microscopy (SEM; Model JSM-5600, JEOL Ltd., Tokyo, Japan). In order to evaluate the pore structure, the obtained porous Si 3 N 4 ceramics were also infiltrated with an epoxy resin and analyzed by SEM using backscattered electron mode (BSE). 3 Results and discussion Fig. 1 shows the typical pore structures of the obtained porous BAS/Si 3 N 4 composites with the different PAM concentrations from 0 to 2 wt.%. With the increase in the amount of PAM, the pore size decreased and the pore structure changed from the lamellar pores with dendrites ( Fig. 1 b) to columnar ones without dendritic pores. In freeze casting, the pore structure is a replica of the morphology of the ice crystal. During the solidification of the suspension, a liquid film should exist between the ice front and the particle, which acts as the transmission medium of water molecules for the ice growth [13] . Fig. 2 demonstrates that a particle with radius r is moving ahead of a solid–liquid interface advanced with a velocity of υ . A particle is either engulfed or ejected by the ice fronts depends on the two competitive forces acting on the suspended particles [17] , i.e. the attractive force ( F η ) and the repulsive force ( F σ ). The attractive force resulted from viscous drag would push the particles to the solid and favor entrapment. It can be given by the following equation [18] : (1) F η = 6 π η υ r 2 d where η is the solution viscosity, r is the particle radius, d is the distance between particle and solid–liquid interface and υ is the velocity of the solidification front. The repulsive force would hinder particle engulfment, which is caused by the different interfacial free energy among solid-particle ( σ sp ), liquid-particle ( σ lp ) and solid–liquid ( σ sl ) interface and can be defined in Eq. (2) [19] : (2) F σ = 2 π r Δ σ 0 a 0 d n (3) Δ σ 0 = σ sp − ( σ sl + σ lp ) where a 0 is the average intermolecular distance in the film, n is an exponent and Δ σ 0 is the balance of the surface free energies at the solid/solution/particle boundary. When F η ≥ F σ , the particles will be pushed to the solid–liquid interface, and then the distance of the gap may be not sufficient to keep the ice crystal surface growing and hence resulting in the particle engulfment. In the obtained porous composites, the main parallel channels from the bottom to the upper surface of the slurry were formed in all three samples, indicating that the particles are ejected by the main ice front owing to F σ ≥ F η . During the growth of the ice crystals, the ceramic particles together with PAM were expelled from the freezing front and thus increasing the solute concentration around the existed main ice crystals, which would increase the attractive force of F η due to the increased solution viscosity. Incorporation of high amount of PAM could further increase the solution viscosity and the attractive force, which eventually would result that the attractive force exceeds the repulsive force. Therefore, the corresponding liquid film (between the ice front and the particle) thickness in the sample with a high PAM content would not be enough to allow the necessary flow of molecules to keep the crystal growing and hence preventing the branch ice crystals growing from the main ice crystals. The porosities analyzed by mercury porosimetry were almost the same as those calculated from the dimensions and weights of the samples, as shown in Table 1 . It reveals that most of the pores in the resultant porous composites are open. The effect of the PAM content on the pore size distribution is shown in Fig. 3 . All the three samples exhibit a bimodal pore size distribution. The small pores below 1 μm correspond to the pores formed by the branch ice crystals and the big ones above 1 μm around to the main channels. With increasing the PAM content from 0 to 2 wt.%, the peak below 1 μm tends to disappear, which is due to the disappearance of branch ice crystals. It is known that the increase in ice crystal size is caused mainly by Ostwald ripening. Ostwald ripening is a well-known phenomenon in which the larger crystals increase in size at the expense of the smaller crystals. In this study, Ostwald ripening is restricted by the mass transport of water. The concentrated solution decreases the mass transfer coefficient of water molecules, thus retard the growth of ice crystals. So when PAM was added to the suspension, small ice crystal was obtained and thus decreasing the pore size, as shown in Fig. 1 . The ice front velocity parallel to the crystallographic c axis (the direction perpendicular to the lamellar ice crystals corresponds to the limited growth direction of ice crystals) is 10 2 –10 3 times lower than the velocity perpendicular to this axis [20] , so as the PAM content increased, the reduction of ice crystal size in the direction of perpendicular to c axis is more evident than parallel to c axis direction, and the pore morphology change from lamellar to columnar. Fig. 4 shows the cross-sectional SEM micrographs of the obtained porous Si 3 N 4 composites, indicating in situ formation of a dense layer at the bottom of the samples with PAM ( Fig. 4 b and c). During freeze casting, initial freezing is very fast and then the velocity of the liquid front decreases rapidly until it reaches a steady state with an approximate constant value. Consequently, the ceramic particles are entrapped by the ice front in the first frozen zone. During sintering, the necks formed by the entrapped particles would favor the atomic transport and form a dense layer at the bottom of the samples, as shown in Fig. 4 b and c. Fig. 4 also reveals that the thickness of the dense layer increases as increasing the PAM content. It is due to the increase of the attractive force for the sample with increased PAM amount. With increasing PAM content from 1 to 2 wt.%, a longer distance is needed for decreasing the freezing velocity to a steady state where F σ ≥ F η , and thus resulting in the increase in the thickness of the dense layer. The influences of PAM additions on the α - to β -Si 3 N 4 transformation and porosity of the investigated porous composites are shown in Table 1 . It is clear that increasing the PAM content obviously reduces the amount of α - to β -Si 3 N 4 transformation and slightly increases the porosity of the obtained samples. It is accepted that the densification and the growth of β -Si 3 N 4 grains in Si 3 N 4 -based ceramics is generally via a liquid phase, through the process of solution–diffusion–reprecipitation. For the samples with PAM, during the growth of the ice crystals, Si 3 N 4 particles together with PAM were expelled from the freezing front. During sintering, the removal of PAM would leave pores on porous Si 3 N 4 walls, and thus retard the formation of necks among particles and impede atomic transport for α - to β -Si 3 N 4 transformation by solution–diffusion–precipitation process. The typical microstructures of the obtained porous composites prepared with different PAM content are shown in Fig. 5 . The aspect ratio of β -Si 3 N 4 grains decreases obviously with increasing the PAM content, indicating that the incorporation of PAM inhibits the anisotropic growth of β -Si 3 N 4 grains. It may be attributed to their different pore structures. In this study, the obtained composites are highly porous, so vapor phase transport cannot be discounted as a contribution to anisotropic β -Si 3 N 4 growth [21] . As stated above, although the incorporation of PAM has no great influence on the porosity of the resultant materials, but significantly changed their pore structures, as shown in Fig. 1 . For the material without PAM addition, although the porosity is slightly lower than that of the materials with PAM, the porosity is highly interconnected with dendritic pores, which would favor the anisotropic β -Si 3 N 4 grain growth through vapor phase transport. In contrast, the addition of PAM obviously changes the pore structure from the lamellar pores with dendrites to columnar ones without dendritic pores, and thus resulting in the hindrance of β -Si 3 N 4 grain growth due to the reduction in vapor phase transport. The flexural strength and fracture toughness of the porous Si 3 N 4 are shown in Table 1 . It reveals that the mechanical properties of obtained porous composites are strongly affected by the microstructure, including pore structure and the morphology of β -Si 3 N 4 grains. Lamellar pores with dentrites in the sample are easy to cause the damage of the obtained porous ceramics. Elongated β -Si 3 N 4 grains and suitable interfacial bonding strength would improve the mechanical properties of the samples. As shown in Table 1 , the sample with no PAM shows the lowest flexural strength and fracture toughness, although it possesses the highest β -Si 3 N 4 content of 77.3% and more elongated microstructure. It may be attributed to its pore character, i.e., lamellar pores with obvious dendrites ( Fig. 1 a and b). The lamellar pores are easier to be torn than the columnar ones. Additionally, the main lamellar channels are interconnected by the dendritic pores, so the rapture might occur along dendritic pores and decrease the mechanical properties. In contrast, the sample with 2 wt.%PAM exhibits increased strength and toughness, although both β -Si 3 N 4 content (38.9%) and the aspect ratio of β -Si 3 N 4 grains are much lower than that of the sample without PAM. It is undoubtedly due to its different pore structure (columnar pores without dendrites, as shown in Fig. 1 e and f). The sample with 1 wt.%PAM exhibits the highest strength and toughness in the resultant porous materials. It is attributed to its columnar pore structure with less dendritic pores and higher aspect ratio of β -Si 3 N 4 grains. 4 Summary Porous BAS/Si 3 N 4 composites with unidirectional aligned channels were fabricated by freeze casting using liquid N 2 as refrigerant. The incorporation of PAM has strong influence on the pore structure, morphology of β -Si 3 N 4 grains and mechanical properties of the obtained porous β -Si 3 N 4 composites. Increasing PAM content in the slurries changed the pore structure from lamellar pores with dendrites to columnar ones without dendrites, decreased the pore size and increased the dense layer thickness. The lamellar pores with dendrites could favor the α → β -Si 3 N 4 transformation, but were detrimental to the mechanical properties. The composite with 1 wt.%PAM possesses the highest strength and toughness in the obtained porous composites. It is attributed to its columnar pore structure with less dendritic pores and higher aspect ratio of β -Si 3 N 4 grains. Acknowledgment This work was supported by National Natural Science Foundation of China (Grant Nos. 90716022 and 50632020 ). References [1] U. Soltmann H. Böttcher D. Koch G. Grathwohl Mater. Lett. 57 2003 2861 2865 [2] J.S. Lee S.H. Lee S.C. Choi J. Alloys Compd. 467 2009 543 549 [3] F.F. Lange Int. Metal. Rev. 1 1980 1 20 [4] M.J. Hoffmann Mater. Res. Bull. 10 1995 28 32 [5] A. Diaz S. Hampshire J. Eur. Ceram. Soc. 24 2004 413 419 [6] J.F. Yang Z.Y. Deng T. Ohji J. Eur. Ceram. Soc. 23 2003 371 378 [7] C. Kawai A. Yamakawa J. Am. Ceram. Soc. 80 1997 2705 2708 [8] J.F. Yang G.J. Zhang T. Ohji J. Mater. Res. 16 2001 1916 1918 [9] K. Ishizaki S. Komarneni M. Nanko Porous Materials—Process Technology and Applications 1998 Kluwer Academic Publishers Norwell pp. 1–11 [10] T. Yokota Chem. Eng. Jpn. 42 1997 845 851 [11] T. Fukasawa Z.-Y. Deng M. Ando T. Ohji Y. Goto J. Mater. Sci. 36 2001 2523 2527 [12] S. Deville E. Saiz A.P. Tomsia Acta Mater. 55 2007 1965 1974 [13] L. Ren Y.P. Zeng D. Jiang J. Am. Ceram. Soc. 90 2007 3001 3004 [14] F. Ye L.M. Liu H.J. Zhang B.S. Wen J. Alloys Compd. 493 2010 272 275 [15] F. Ye J.C. Gu Y. Zhou M. Iwasa J. Eur. Ceram. Soc. 23 2003 2203 2209 [16] S.W. Quander A. Bandyopadhyay P.B. Aswath J. Mater. Sci. 32 1997 2021 2029 [17] G. Lipp C. Körber G. Rau J. Cryst. Growth 99 1990 206 210 [18] G.F. Bolling J. Cissé J. Cryst. Growth 10 1971 56 66 [19] D.R. Uhlmann B. Chalmers K.A. Jackson J. Appl. Phys. 35 1964 2986 2993 [20] V.F. Petrenko R.W. Whitworth Physics of Ice 2002 Oxford University Press [21] C. Kawai A. Yamakawa Ceram. Int. 24 1998 135 138